The summary you run before anything else. What each column tells you, the two comparisons that catch most data problems in seconds, and how to report descriptives in APA 7.
Quanta produces one row per numeric field with N, missing, mean, median, SD, minimum and maximum, and reports how many fields are included.
That is a deliberately short list. It contains the four things that catch most data problems and leaves the rest for the analyses that need them.
Run it before every analysis, not once at the start of a project. Descriptives are cheap and the errors they catch are expensive.
Compare the minimum and maximum against your instrument. A five-point item should run 1 to 5. A 6 means a coding error. A 0 means a missing-value code that was never declared, and it will be silently averaged into your mean. This single check catches more real errors than any assumption test.
Compare the mean against the median. When they sit close together, the distribution is roughly symmetric. When the mean is pulled well away from the median, something is dragging it: skew, outliers, or a stray code. On a 1-to-5 item a mean of 2.92 with a median of 3.00 is unremarkable; a mean of 4.6 with a median of 3.0 is a signal.
Read the missing column per field, not as a total. Fields with concentrated missingness behave differently from fields with scattered missingness, and listwise deletion in a later analysis will remove any case missing any field. A dataset with 250 complete cases per variable and one with 250 cases each missing a different variable produce very different analysis samples.
Check the SD is not zero. A field with no variance cannot correlate with anything, cannot predict anything, and will silently break models that include it.
The count tells you how much data is absent. It does not tell you whether the people who did not answer differ from the people who did.
If non-response is concentrated in one group, then every analysis using that field describes the remaining group rather than your sample. That is a question about representation, not about statistics, and it cannot be repaired downstream by a technique.
Quanta treats this as a first-class question rather than a footnote: whether missingness falls evenly across groups is part of Data Quality, not something you have to think to check.
Open Descriptives & Explore and select the fields you want summarized. Quanta reports the count of included fields so you can confirm nothing was left out.
Read min and max first, then mean against median, then the missing column. That order surfaces data errors before you start interpreting anything.
Use the action bar to add the table to a report, copy it, or copy the APA text. Every result carries the ReliCheck Intelligence card for a plain-language explanation computed on your Mac.
Report means and standard deviations to two decimal places, with the sample size and the scale range.
Responses to the self-awareness items (1 to 5 scale) averaged between 2.89 and 3.11 (SD range 1.25 to 1.31), with no missing data (N = 250).
Give the scale range. A mean of 2.92 means nothing without knowing whether the scale runs 1 to 5 or 1 to 7.
Report missing data and how it was handled. Saying nothing implies there was none, which is a claim you are making whether you intend it or not.
For skewed distributions report the median and an interquartile range instead of, or alongside, the mean. A mean that misrepresents the centre is worse than no summary at all.
Every result carries a ReliCheck Intelligence card with one action: explain this result in plain language. It runs entirely on your Mac with Apple Intelligence. No upload, no API key, nothing to configure. The model only explains the result Quanta has already computed and validated. It never computes a statistic, and it never applies a decision on your behalf.
From the action bar you can add the result to a report, copy the table, or copy the APA text. Exported reports carry the APA table, its note, and a caution specific to that analysis beneath it.
Quanta's univariate summaries are validated against the NIST Statistical Reference Datasets, matching certified means to 12 or more digits and standard deviations to about 8 using a cancellation-safe algorithm. NIST publishes these datasets because naive implementations fail them. The engine is pure Swift with no R or Python at runtime. Full record on the validation page.
Descriptives & Explore is part of every Quanta subscription, $9.99 a month or $79.99 a year after the 30-day trial. One price for everyone. See pricing.
The analysis, the report builder, and the on-device explanations all work with the network off. Your dataset is never uploaded.
Normality tests Shapiro-Wilk and Lilliefors · Demographics tables APA frequencies for categorical fields · Stem-and-leaf seeing the shape of a distribution · the complete list on the analyses page.